More Options

The Poor, Misunderstood Placebo

Steven Novella

Volume 34.6, November/December 2010

Understanding placebo effects is critical to making sense of medical research and ever-expanding health claims within an increasingly unregulated market.

A recent study looking into the effects of acupuncture on relieving back pain was widely reported in the media as finding that "acupuncture works, even fake acupuncture." Behind the headlines, the authors were more circumspect in the paper itself, concluding:

Although acupuncture was found effective for chronic low back pain, tailoring needling sites to each patient and penetration of the skin appear to be unimportant in eliciting therapeutic benefits. These findings raise questions about acupunctures purported mechanisms of action. It remains unclear whether acupuncture or our simulated method of acupuncture provide physiologically important stimulation or represent placebo or nonspecific effects. (Cherkin et al. 2009)

The authors compared acupuncture to placebo acupuncture (creating the sensation of acupuncture with toothpicks that do not penetrate the skin) and found no difference. When a drug is compared with a placebo and there is no difference in the response, the standard conclusion is that the drug has no effect--it does not work. But that logic is being turned on its head by what is being called "placebo medicine," in which the placebo effect is seen as a real, valuable, and desired outcome of patient treatment (Novella 2009).

Placebo effects (plural), however, are varied and complex. For the most part they are not a biological response to the expectation of benefit, which is what most people assume. Understanding placebo effects is critical to making sense of medical research and ever-expanding health claims within an increasingly unregulated market.

Operational Definition of Placebo

In the context of research, the "placebo effect" has a very specific operational definition: it is the treatment effect measured in the placebo arm of a clinical trial, which includes those subjects who have received a fake or inert treatment. If a trial is rigorously designed, placebo effects should include everything other than a physiological response to a biologically active treatment. Therefore, we can subtract placebo-effect findings from the treatment group, who will display treatment effects plus placebo effects, and we are left with a measurement of the treatment effect alone.

This simple but effective logic is the cornerstone of medical research. It is necessary because there are a variety of effects that can create the false impression that a treatment is working even when it isn't. It is a mistake to assume that the only relevant false impression is a "mind over matter" effect resulting from belief in the treatment. This is not the case.

Placebo Effects

Placebo effects fall into several categories: illusions of observation, bias, nonspecific effects, and physiological effects. Much of what is measured as a placebo effect is, in fact, simply an illusion of the process of observation. These illusory effects include regression to the mean, which is a statistical phenomenon that includes extreme symptoms becoming less extreme as a matter of course. For any variable symptom, periods of time when symptoms are at their worst are likely, by chance alone, to be followed by a return to more average symptoms.

Other artifacts include the biases of the researchers and the subjects. Researchers want their interventions to work and may therefore bias their assessments to be more positive. Subjects want to receive an effective treatment and to meet the expectations of the researcher. They want to justify their risk, expense (even if its just time), and their decision to receive a treatment or enter a trial. Conditioning, in which one associates a treatment ritual with feeling better, is another related placebo effect.

There are also a number of nonspecific effects, such as the well-documented observer effect (also called the Hawthorne effect), in which the very fact of being observed in a clinical trial results in a change in behavior and reporting (McCarney et al. 2007). People are more likely to be compliant with treatment, take better care of themselves, and get regular medical attention as part of a trial. Related to this is the cheerleader effect: for any functional assessment, people will tend to try harder if they are being encouraged, if they feel they should be doing better, or if they have hope that the treatment is working.

And finally there are real physiological effects resulting from the ritual of treatment. For example, treatment may involve relaxation or simply taking a break from your otherwise hectic daily routine. Believing one is being treated may reduce anxiety about the illness or symptoms, which in turn may reduce sympathetic activity, reduce blood pressure and strain on the heart, and reduce the levels of stress hormones. Hands-on treatments have the benefit of human contact, which improves mood and provides an overall feeling of well-being.

The perception of pain in particular is subject to these nonspecific effects, such as when an improved mood reduces the perception of pain. In addition, conditioning, expectation, and nonspecific benefits may actually cause the release of natural endorphins that reduce pain transmission (Benedetti 2007) or the release of dopamine in the reward centers of the brain (de la Fuente-Fernández and Stoessl 2004).

Breaking It Down

Given this more thorough understanding of placebo effects, it is not reasonable to assume that the measured placebo effect in a clinical trial is mostly or entirely a real "mind over matter" health benefit. Instead, the placebo effect may consist mostly or entirely of illusion and bias. One might ask, "Which kinds of effects are contributing to the measured placebo effect of specific treatments?" The answer is that it depends on what is being treated.

For example, a study of placebo effects in the treatment of irritable bowel syndrome (Kaptchuk et al. 2008) found a substantial placebo effect in place with the use of placebo acupuncture. The same study also found that the enhanced placebo group--members of which received placebo acupuncture with enhanced interaction between the therapist and the subject--reported added "warmth, attention, and confidence." After three weeks, the waiting-list group (those who received no treatment--not even placebo) had about a thirty-point drop on the symptom severity scale, with almost 30 percent of patients reporting adequate relief; the treatment group (those who received placebo acupuncture) had a forty-two-point drop, with 44 percent of patients reporting relief; the augmented group (those who received placebo acupuncture plus enhanced interaction with the therapist) reported a drop of over eighty points, with over 60 percent of patients reporting relief.

It's very interesting that the group that received no intervention, not even a placebo, still had a 30 percent response. This response is likely entirely due to observational artifacts (Hawthorne effect, etc.). The placebo intervention also led to an improved response--in this case, expectation and conditioning might be having an effect. The enhanced intervention group showed the strongest effect, likely representing an increase in the nonspecific benefits of a positive therapeutic relationship.

Hróbjartsson and Gøtzsche have been studying placebo effects for years. They recently reviewed clinical studies that contain a no-treatment arm as a way of measuring placebo effects. They conclude:

We did not find that placebo interventions have important clinical effects in general. However, in certain settings placebo interventions can influence patient-reported outcomes, especially pain and nausea, though it is difficult to distinguish patient-reported effects of placebo from biased reporting. The effect on pain varied, even among trials with low risk of bias, from negligible to clinically important. Variations in the effect of placebo were partly explained by variations in how trials were conducted and how patients were informed. (Hróbjartsson and Gøtzsche 2010)

Let's break this down a bit. First, Hróbjartsson and Gøtzsche found that when you look at any objective or clinically important outcome--the kinds of things that would indicate a real biological effect--there is no discernible placebo effect. That is, there is no "mind over matter" self-healing that can be attributed to the placebo effect.

What the authors found is also most compatible with the hypothesis that placebo effects, as measured in clinical trials, are mostly due to bias. Specifically, significant placebo effects were found only for subjectively reported symptoms. Further, the size of this effect varied widely among trials.

Conclusion

Increasingly, placebo effects are being used to justify the use of ineffective and even inert treatments, with the assumption that "the" placebo effect is a true healing effect. What the research indicates, however, is that there are many placebo effects, and they are mostly bias and illusion--not real effects. There are also nonspecific effects that are likely valuable, but these effects can mostly be categorized as stress reduction and improvement in mood through attention and encouragement. It should be remembered that any placebo effect worth having will also accompany a legitimate treatment that actually works. On close inspection, placebo effects are not a justification for substituting hocus-pocus for real medicine.

References

Benedetti, F. 2007. Placebo and endogenous mechanisms of analgesia. Handbook of Experimental Pharmacology 177: 393-413.

Cherkin, D.C., K.J. Sherman, A.L. Avins, J.H. Erro, L. Ichikawa, W.E. Barlow, K. Delaney, et al. 2009. A randomized trial comparing acupuncture, simulated acupuncture, and usual care for chronic low back pain. Archives of Internal Medicine 169(9): 858-66.

De la Fuente-Fernández, R., and A.J. Stoessl. 2004. The biochemical bases of the placebo effect. Science and Engineering Ethics 10(1): 143-50.

Hróbjartsson, A., and P.C. Gøtzsche. 2010. Placebo interventions for all clinical conditions. Cochrane Database Systematic Review (1): CD003974.

Kaptchuk, T.J., J.M. Kelley, L.A. Conboy, R.B. Davis, C.E. Kerr, E.E. Jacobson, I. Kirsch, et al. 2008. Components of placebo effect: Randomised controlled trial in patients with irritable bowel syndrome. British Medical Journal 336(7651): 999-1003.

McCarney, R., J. Warner, S. Iliffe, R. van Haselen, M. Griffin, and P. Fisher. 2007. The Hawthorne Effect: A randomised, controlled trial. BioMed Central Medical Research Methodology 7: 30.

Novella, S.P. 2009. The rise of placebo medicine. Available online at www.sciencebasedmedicine.org/?p=672.

Steven Novella

Steven Novella, MD, is an assistant professor of neurology at Yale University School of Medicine. He is the host of the Skeptics’ Guide to the Universe podcast, author of the NeuroLogica blog, executive editor of the Science-Based Medicine blog, and president of The New England Skeptical Society.